How ultralytics/notebooks works
Project type and purpose
• This is an “example‐and‐documentation” repository. It provides a set of Jupyter notebooks showcasing Ultralytics YOLO workflows, plus tools for generating and formatting documentation tables in the README. There is no monolithic “application” here—rather a content library, CI pipelines, and helper scripts for docs.
File structure
• Top level:
• .github/workflows – CI pipelines (formatting, table generation)
• docs – configuration data (notebooks-data.yml) and a Python script (update-readme-table.py)
• notebooks – dozens of .ipynb example notebooks
• pyproject.toml – Python project metadata & dependencies
• README.md – landing page with table of notebooks
• External integrations: Google Colab, Kaggle, SageMaker badges embedded in README per notebook; links to YouTube, Ultralytics HUB, blogs, arXiv, etc.
README for additional insights
• The README is auto-generated by docs/update-readme-table.py reading docs/notebooks-data.yml.
• CI workflow table.yml runs on push to regenerate the table.
• format.yml enforces code & markdown formatting, probably via black/isort/pre-commit.
Elements
• Clear labels: “notebooks/”, “docs/”, “.github/workflows/”, “README.md”, “pyproject.toml”, external services
• docs/notebooks-data.yml → update-readme-table.py → README.md
• Push → GitHub Actions workflows → formatting & docs generation
• README badges → launch notebooks on Colab / Kaggle / SageMaker
• Blue rectangles for static content (notebooks, YAML)
• Green rectangles for scripts (update-readme-table.py)
• Orange rectangles for CI workflows
• Gray clouds for external services
Detailed capture of essential elements
• Dual CI workflows side by side, each pointing to the files they act on, and to the README as their output.
• External integrations as end‐user entry points triggered by badges in the README.